New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Comments
If the host doesn't support ISOs, I usually just use netboot.xyz, either through iPXE or grub-imageboot depending on if the host exposes iPXE during system boot. If iPXE works, use it (https://netboot.xyz/docs/booting/ipxe), otherwise boot to netboot.xyz via GRUB (https://netboot.xyz/docs/booting/grub).
What I'm wondering is what HostHatch's control panel considers a "reinstall" for the purpose of differentiating "legacy" from "non-legacy" VPSes.
It looks like reinstalling actually reduces the size of the disk
I'm unable to restore my Clonezilla disk backup image because the disk is now too small:
its been confirmed on OGF that clicking on reinstall will:
So now I have no way to restore my VPS from backup, since it physically won't fit on the disk (the disk was actually pretty close to full)? I wonder if there's a way to un-migrate and switch back to legacy. I'll ask their support. I guess it's good I caught this on a test VPS before deciding to do it on my important ones... Those can just stay as legacy ones for now.
Pretty much every other VPS provider uses GiB for disk space though, even if they say "GB".
to answer my own question, RAM is shown as GiB in new panel.
Hosthatch described the procedure on the OGF like you said. They're destroying and recreating your VPS so the machine can be used for all new features (private networking, snapshots etc.). This also makes the units converting from GB to GiB. @hosthatch wanted to add a note at their webpanel explaining this move as this has lead to confusion in several places now.
Same with mine
Now my NVMe vps in LA cannot boot and cannot be reinstalled
On panel showing offline
It is unused and broken
Private networking is definitely available on the legacy VMs, you just need to opt-in in the control panel (at least in Los Angeles). I'm using it and it works well. In fact the same private network includes both legacy and new VMs
Had the same problem, turned out to be a browser extension, but couldn't figure out which. Privacy Badger and uMatrix are supposed to be my only blockers in Chrome, but with both disabled it still didn't work.
Started Edge from command line without extensions and it worked.
Typically with other providers, a reinstall would get the same IP assigned when the VM is recreated. It looks like in the new panel, you get new IP's (not loving this) and MAC's. So I wouldn't expect you manually doing an ISO reinstall would have anything to do with the panel "Reinstall".
In advanced, you can skip checking the final size, but you really need to have less data than the final partition size.
But this is why I hate clonezilla and don't understand why something else hasn't replaced it. I think even free Veeam agent will restore Linux to smaller destinations.
How do I check the size of the image? None of the text files in the Clonezilla image directory seem to list have the amount of space actually used on the partition. The reason I'm worried about it is because I remember the used space being pretty close to the limit.
If you remember being less than 10% free, you're fucked.
I think you might need to upload it somewhere with space and extract it to see the real size, but being an image, its probably padded out and not to the actual data size.
You don't have a .compact file there? Usually it has an fdisk style partition list there.
Disk /dev/sda: 250GB
Sector size (logical/physical): 512B/512B
Partition Table: msdos
Disk Flags:
Number Start End Size Type File system Flags
1 1049kB 106MB 105MB primary ntfs boot
2 106MB 250GB 250GB primary ntfs
That's just the partition table. It doesn't say how much space was actually used in each partition. What I need to know is if all the files that were previously on the partition would still fit on the new, smaller-sized one.
Yes but then you can just uncompress it with gzip, and loop mount and fsck later.
That's just the same as with dd but with gzip:
cat image.gz.* | gzip -d > image.img
Oh yeah, loop mount is a good idea! I used zstd rather than gzip but it'd be similar.
In any case, it doesn't really matter right now since HostHatch's newer control panel is currently down. Maybe I'll try it over the weekend. This particular VPS isn't overly important, which is why I used it for testing.
As long as you get the data back, even on another VPS or a local machine, that's why I try to
give some advices how to still recover and fix it with fsck later.
Wonder what made them do such odd "migration" in the first place, an introduction of some "private
network feature" has nothing to do with it, usually must be a solid reason to cause that amount
of downtime.
I imagine the technical maintenance debt became too burdensome and thought they could do better with their own panel. There needs to be a one-time painful migration in any event.
my guess they wanna avoid getting Virmach-ed, as in fucked by SolusVM*WHMCS in all directions on such a large scale
so now customers are being Hosthatch-ed instead.
This is the reason I don't deal with such summerhosts for anything important.
I don't mean particularly the ones you mentioned, there are many well respected
providers here with skillful staff. But still, I just can't imagine such things happening
with the big players like Vultr/DO.
This is not correct. Please clarify how you came to this conclusion. And if you do not have any facts to back this up with, please either edit your post, or make another post in bold so people don't open tickets with us asking whether/why we are changing IPs at random. Spreading false information is not appreciated.
What amount of downtime? A reboot and about 10 minutes? Have you ever had maintenance on any of the existing providers you use? All the "downtime" complaints involved here are one of this:
In both cases, the assumption is, that your production is not being affected, and a few hours of normal support queue wait is acceptable. The same would be the case with any of the bigger providers you mentioned.
I know there is probably a bunch of mis/half information involved here, but I am happy to answer any actual questions. This is why I rarely come around these forums here, since creating drama seems to be a higher mission than actual facts and complaints.
You've never had a 10 minute downtime and a reboot on Vultr/DO/Linode, or is that something you just made up? I have several maintenance emails from them that I would be happy to share with you since I clearly seem to have not been as lucky as you. I have also had much larger downtimes with those providers. But still, it sounds cool to say "this wouldn't happen to X because they are big"
We've spent months on making this as painless as possible for end-users, and are working very hard on the process of actually doing it and making sure any complaints are addressed as soon as possible.
I'm not trying to put your business in a bad light, I just comment on the situation I see.
A few complaints from different users, some of them are clearly with good Linux administration
skills are talking about instances that can't boot, shrinked partitions, inability to restore dd backups and IP changes - this is way more serious than a "10 minute downtime" in my opinion. I never
had anything described in this thread with the "big" providers, ever. That's why I'm lucky penguin
Indeed, and that is my problem with the half information people choose to sometimes share to make it look like a bigger issue than it is. I am not saying we are handling this perfectly, but hitting someone on the arm and murdering someone are two different severities of a crime. Please allow me to clarify all the points you mentioned:
The only real complaint I can see is the GiB to GB conversion **after ** a reinstall (not before, the control panel migration just allows the same old VM to be managed through the new panel, it does not shrink any partitions - that would be stupid of us and would render most VMs useless). However, we've heard the feedback and added a note on the reinstall page that clarifies this information in advance.
The only downtime involved here from our end is a shut down to the VM, updates to the host node, and a boot to the VM. 10-15 minutes. That is it.
Most of the longer downtime that people mention here is either one of those cases I explained above.
No IP addresses are changed, at all. This is just wrong.
Hopefully my above comment clarifies it for you, since we also didn't have the things described in this thread happen to us, not in the doomsday way they are being presented anyway.
Can you check
My LA NVMe vps cannot controlled via panel (new and old panel)
It always show stopped, cannot bootable nor be reinstalled
Looks like it is broken
server = 54406
Already made a ticket but since you online here i can make a report instead
Just had four servers migrated in Stockholm without a hatch hitch. Liking the new panel.
The routed IPv6 and the new panel are very prem
I had a 1 TB storage server involucrated in Chicago and it was remade on the new panel (which got a new IP, as expected since old server was online. Thats fine). That didn't work, I couldn't login with the root password in panel. I did a reinstall in the panel and still couldn't login. It got a new IP. I opened a ticket and they remade the VM. It got a new IP.
I don't know if I have spreadsheet tracking and/or history saves to give you the IP's, I just recall making several DNS and spreadsheet changes. It wasn't an issue in this case I was just concerned for the future.
But if you're saying that doing a reinstall in the panel shouldn't assign a new IP, that's great and I'll just chalk it up to teething issues on the new panel.
Also, doing a specific action (reinstall) is not a random thing nor was there anything remotely implying such happened, so you can chill out, you're overreacting.
Lastly, you can only edit posts for one hour and there's little to no chance you'd respond and I could respond within that time, so refrain from making such requests or you'll look irrational. It's fair to ask for clarification. It also makes sense to just quote and refer back to previous posts instead of editing conversations and confusing the thread with removed discussion.
Not sure what any of this has to do with the panel migration that's being discussed here, but please understand that adding unconfirmed assumptions only adds confusion and only makes things worse for everyone else.
In a completely unrelated event to this panel migration, we created a new VM for you. It did not work, so we deleted that VM and created a new one for you, which had a new IP address. This is what I understand from your latest post.
In any case, a reinstall will never give you a new IP address without intervention on our side, it is just not possible.